On Mon, Feb 24, 2020 at 12:08 PM Lukas Ruzicka <lruzicka(a)redhat.com> wrote:
Hello, I have been doing the KDE "all applications must
work" test for
more than two years already and I can
tell you, that some of the KDE pre-installed applications are of low
quality, if not broken already. Last time I checked, there were over 60
applications listed in Menu where there are like 30 in Gnome. Some
applications (browsers) are tripled, some are doubled. Others are connected
with a certain use case, e.g. converting one data format into another,
therefore useless for people without that use case.
If "all applications" must work, than I need to start every application
and test the "basic functionality". Nowhere is written exactly, what basic
functionality is, so I must test what I think needs to be tested "in bona
fide." Many KDE applications generally work, but they have flaws in some
parts that can or do not need to be basic functionality, depending on
opinions.
In case of a reported bug, the readiness of KDE developers to fix it
usually is lower when compared to Gnome, so sometimes there are bugs that
will never be fixed, because they are not as severe as to block the release
completely, so they get ship over and over again.
The proposal could actually improve the situation, because important apps
would be tested thoroughly and not just for the basic functionality.
Lukas described it all nicely, but in this particular place I have to
clarify. The proposed criterion still requires just "basic functionality"
of covered apps. If we can use some of the saved time to test those covered
applications more thoroughly is a different question, and it's possible,
but I don't want to promise anything. Another possible improvement is that
we might be able to automate testing of some of those applications.
I want to make one additional point. I could have proposed that the
criterion applied to all desktops. From QA point, it's less work for us,
that should make us happy, right? But I didn't - because I'm trying to find
some balance between our responsibilities and the end-user benefits. And
it's a difficult balance to find. If we can't find it, that easy solution
is always there. If you have opinions on where to draw the line and how to
improve the situation in a different way, please tell us.